An automatically controlled robot arm

Abstract

The application relates to an automatically controlled mechanical arm, comprising: a mechanical arm body with an execution end for executing actions; a clamping mechanism arranged at the end of the mechanical arm body and used for executing clamping actions; a detection module arranged at the end of the mechanical arm body and used for detecting a work target and a surrounding environment; a fine adjustment platform, wherein the mechanical arm body is arranged on the fine adjustment platform and used for secondarily adjusting the position of the mechanical arm body; and a control module in signal connection with the mechanical arm body, the clamping mechanism, the detection module and the fine adjustment platform respectively; the control module controls the detection module to scan the work target and the surrounding environment and obtain scanning information according to a work task. The application can realize automatic work control of the mechanical arm, does not need close control by a person, can be applied to scenes where a person cannot enter or is in danger, and has better scene applicability.

Description

Technical Field

[0001] This application relates to the technical field of robotic arms, and more specifically to an automatically controlled robotic arm. Background Technology

[0002] A robotic arm is a complex system characterized by high precision, multiple inputs and multiple outputs, high nonlinearity, and strong coupling. Due to its unique operational flexibility, it has been widely used in fields such as industrial assembly and safety and explosion protection.

[0003] Existing robotic arms typically require control via a teach pendant or controller. Specifically, the operator holds the controller within a certain range of the robotic arm and controls the robotic arm to perform corresponding actions according to the needs of the task.

[0004] This control method requires the operator to be within a certain range of the robotic arm body. In areas where operators cannot enter, such as dangerous areas or radiation areas that could easily cause harm to personnel, the above-mentioned robotic arm cannot be used.

[0005] The second type of robotic arm performs repetitive actions by programming its prescribed movements, such as in equipment assembly. However, this method has highly targeted motion paths, which are relatively singular or limited to a few pre-defined ones, making it difficult to efficiently adjust them according to actual conditions.

[0006] The third type of robotic arm is manually adjustable, such as using a magnetic base, and is mainly used in precision measurement. This type of robotic arm is difficult to precisely control in terms of fine-tuning, requiring a high level of skill from the operator.

[0007] Therefore, there is an urgent need to design a robotic arm that can achieve automated control and solve the technical problems existing in the above-mentioned technologies. Summary of the Invention

[0008] To address the problems existing in the prior art, this application aims to provide an automatically controlled robotic arm. This application enables automated operation control of the robotic arm, eliminating the need for close-range manual control, and is applicable to scenarios where personnel cannot access or where such access is dangerous, thus exhibiting better scenario applicability.

[0009] The automatically controlled robotic arm described in this application includes: The robotic arm body has an actuator for performing actions; A gripping mechanism is located at the end of the robotic arm body and is used to perform gripping actions; A detection module, located at the end of the robotic arm, is used to detect the target and the surrounding environment. A fine-tuning platform is provided, on which the robotic arm body is mounted for secondary adjustment of the position of the robotic arm body. The control module is connected to the robotic arm body, gripping mechanism, detection module and fine-tuning platform respectively; The control module controls the detection module to scan the target and surrounding environment to obtain scanning information according to the task. Based on the scanned information, a three-dimensional model of the target task and its surrounding environment is constructed. Based on the task, optimal path planning is performed using the 3D model to obtain the optimal running path for the execution end; Based on the obtained optimal running path, control the execution end to move to the initial predetermined position; Based on the positional deviation between the target position and the initial position, the fine-tuning platform is controlled to move the robotic arm body to the target position, thereby moving the execution end to the target position. The robotic arm body is instructed to perform a predetermined operation.

[0010] Preferably, the gripping mechanism includes a pneumatic gripper or an electric gripper; the detection module includes one or more of an infrared detector, an ultrasonic detector, a lidar, and a 3D scanner.

[0011] Preferably, the fine-tuning platform includes an X-axis walking module, a Y-axis walking module, a Z-axis walking module, a first drive component, a second drive component, and a third drive component. The first drive component is linked with the X-axis walking module, the second drive component is linked with the Y-axis walking module, and the third drive component is linked with the Z-axis walking module.

[0012] Preferably, the automatically controlled robotic arm further includes a displacement sensor, which is disposed at the fine-tuning platform to detect the displacement of the fine-tuning platform.

[0013] Preferably, the automatically controlled robotic arm further includes a control terminal, which is signal-connected to the control module.

[0014] Preferably, the control module controls the detection module to scan the target and its surrounding environment to obtain scanning information according to the task, including the following steps: The control module controls the detection module to scan at a first speed. Perform an initial scan of the target and its surrounding environment to obtain initial scan information; Based on the initial scan information, an initial model of the task target and its surrounding environment is constructed; In the initial model, a specific area containing the target task is selected, and the second scanning speed of the detection module is controlled. A second scan is performed on the specific area to obtain second scan information, wherein the first scan speed... Greater than the second scan speed ; Based on the secondary scan information, the initial model is optimized to obtain a three-dimensional model of the target and its surrounding environment.

[0015] Preferably, the control module controls the detection module to scan the target and its surrounding environment to obtain scanning information according to the task, including: Construct complexity coefficients to characterize the complexity of the task objective and its surrounding environment. , Establish the aforementioned complexity coefficients With the first scan speed Second scan speed The complexity-speed mapping relationship; Based on the initial model, the complexity coefficient of the task objective is calculated. : , , , , , in, The geometric complexity coefficient of the task objective is represented. Indicates the environmental interference coefficient. and These represent the weights of the geometric complexity coefficient and the environmental disturbance coefficient, respectively. The geometric complexity evaluation value represents the objective of the task. The standard value representing the geometric complexity of the task objective; The coefficient representing the rate of change of surface curvature of the target object. The contour regularity coefficient represents the objective of the task. and These represent the weights of the surface curvature change rate coefficient and the contour regularity coefficient, respectively. This represents the environmental disturbance evaluation value for the operational objective. This represents the standard value of environmental disturbance to the operational target. The obstacle density coefficient representing the environment surrounding the target operation. This represents the distance coefficient between the target object and obstacles. and These represent the weights of the obstacle density coefficient and the spacing coefficient, respectively. Based on the complexity coefficient of the stated task objective Based on the complexity-speed mapping relationship, a first scanning speed suitable for the target task is found. Second scan speed .

[0016] Preferably, the optimal running path of the execution end is obtained by performing optimal path planning based on the three-dimensional model according to the task, including the following steps: The coordinate data of the robotic arm body, the fine-tuning platform, and the target are acquired and normalized to obtain a normalized coordinate system. Based on the normalized coordinate system, the relative coordinate relationship between the current execution terminal and the task target is calculated. Based on the relative coordinate relationship and the task, the optimal running path of the execution terminal is planned through a path planning algorithm.

[0017] Preferably, controlling the fine-tuning platform to move the robotic arm body to the target position based on the positional deviation between the target position and the initial position includes the following steps: Obtain the coordinate information of the target location. Obtain the coordinate information of the initially determined position. ; Calculate the positional deviation between the target position and the initial position. : , Control the fine-tuning platform according to the obtained position deviation move.

[0018] Preferably, controlling the movement of the fine-tuning platform further includes: calculating the motion compensation amount of the robotic arm body. Combined with the aforementioned motion compensation amount and positional deviation Controlling the movement of the fine-tuning platform, the motion compensation amount The calculation steps include: A first prediction model is constructed with ambient temperature as input and the X-axis, Y-axis, and Z-axis offsets of the robotic arm body as outputs; a second prediction model is constructed with vibration acceleration as input and the X-axis, Y-axis, and Z-axis offsets of the robotic arm body as outputs. The ambient temperature of the target object and the vibration acceleration of the robotic arm body are collected in real time. The collected ambient temperature is input into the first prediction model to obtain the X-axis temperature error of the robotic arm body. Y-axis temperature error and Z-axis temperature error ; The collected vibration acceleration is input into the second prediction model to obtain the X-axis vibration error of the robotic arm body. Y-axis vibration error and Z-axis vibration error ; The motion compensation amount Represented as: , , in, This indicates the X-axis compensation amount. Indicates the Y-axis compensation amount. This indicates the Z-axis compensation amount.

[0019] The automatic control robotic arm described in this application has the advantage that, by setting up a detection module and a control system, the robotic arm can achieve automated control. The detection module at the end of the robotic arm scans the target, performs optimal path planning based on the scanning results, and then controls the robotic arm to perform the corresponding operation. This achieves automated operation control of the robotic arm without the need for close-range manual control. It is applicable to scenarios where personnel cannot enter or where there is a danger to personnel, and has better scenario applicability.

[0020] This application, by setting up a fine-tuning platform, allows for secondary fine-tuning of the actuator position based on the adjustment of the robotic arm itself, thereby improving the adjustment accuracy of the robotic arm end. Furthermore, this application calculates motion compensation amounts based on the interference caused by ambient temperature and vibration on the robotic arm, and adjusts the position of the robotic arm through the fine-tuning platform based on these motion compensation amounts, improving positioning accuracy and thus enhancing the precision of the work process.

[0021] This application, by setting a mode of initial scanning + secondary precision scanning when scanning the target, can balance scanning accuracy and scanning efficiency, and has better engineering applicability.

[0022] This application normalizes different coordinate systems through a coordinate normalization algorithm, and then realizes the optimal path planning of the robotic arm through a path planning algorithm. This enables the robotic arm to complete automatic path planning and action according to the work target and task, significantly improving the degree of automation. Attached Figure Description

[0023] Figure 1 This is a three-dimensional schematic diagram of an automatically controlled robotic arm as described in this application; Figure 2 This is a side view of an automatically controlled robotic arm as described in this application; Figure 3 This is a schematic diagram of the structure of the execution end described in this application; Figure 4This is a structural block diagram of an automatically controlled robotic arm as described in this application.

[0024] Explanation of reference numerals in the attached drawings: 1-robotic arm body, 2-fine adjustment platform, 21-Z-axis walking module, 22-X-axis walking module, 23-Y-axis walking module, 3-gripping mechanism, 4-detection module, 5-control module, 61-first drive component, 62-second drive component, 63-third drive component, 7-control terminal. Detailed Implementation

[0025] like Figures 1-4 As shown, the automatically controlled robotic arm described in this application includes: The robotic arm body 1 has a six-degree-of-freedom actuator that can move according to control commands.

[0026] Fine-tuning platform 2 includes an X-axis travel module 22, a Y-axis travel module 23, a Z-axis travel module 21, a first drive component 61, a second drive component 62, and a third drive component 63. The first drive component is linked with the X-axis travel module, the second drive component is linked with the Y-axis travel module, and the third drive component is linked with the Z-axis travel module. For example, the first drive component 61, the second drive component 62, and the third drive component 63 are all servo motor + ball screw transmission structures to achieve high-precision secondary fine-tuning.

[0027] A gripping mechanism 3 is disposed at the end of the robotic arm body 1 and is used to perform gripping actions. The gripping mechanism 3 specifically includes pneumatic or electric grippers, or other mechanisms that can be used to grip and fix tools, such as suction cups. The object gripped depends on the task. For example, in measurement tasks, measuring instruments such as dial indicators and thermometers are used, while in disassembly tasks, tools such as pliers are used. This invention does not impose any limitations on this. In actual operation, a suitable gripping mechanism 3 and corresponding tools can be selected according to different tasks.

[0028] A detection module 4, located at the end of the robotic arm body 1, is used to detect the target and its surrounding environment. Specifically, the detection module 4 senses the target and obtains its position coordinates, providing a data source for subsequent path planning. The detection module 4 can be one or more of infrared detection, ultrasonic detection, lidar, or a 3D scanner, used to detect the target and its surrounding environment to generate target position information and environmental information. In a specific embodiment, by reading real-time data from the robotic arm body 1, dynamic information such as the motion state and real-time position information of the robotic arm body 1 can be obtained, such as displacement, velocity, and acceleration information in the X, Y, and Z directions.

[0029] The fine-tuning platform 2, on which the robotic arm body 1 is mounted, is used for secondary adjustment of the position of the robotic arm body 1. The fine-tuning platform 2 adjusts the position of the robotic arm body 1, thereby adjusting the position of the actuator. Its function is to perform a secondary fine-tuning of the position of the robotic arm body 1 based on the initial adjustment. Specifically, based on the position information of the work target and the position information of the robotic arm body 1, the fine-tuning platform 2 adjusts the position of the robotic arm body 1 in the XYZ directions to achieve greater precision. The fine-tuning platform 2 specifically adopts an electric three-axis platform, using a servo motor + ball screw / gear transmission structure to drive the traveling table in the XYZ directions. Combined with the high-precision characteristics of the servo motor and ball screw, the adjustment accuracy can be improved to the micrometer level.

[0030] Furthermore, a displacement sensor is also installed on the fine-tuning platform 2 to accurately obtain the distance traveled by the fine-tuning platform 2.

[0031] The control module 5 is connected to the robotic arm body 1, the gripping mechanism 3, the detection module 4, and the fine-tuning platform 2 respectively. For example, the control module 5 includes a processor, which is an MCU or a PLC. The control module 5 is mainly used to control the actions of each component according to a preset program to complete the automated operation task.

[0032] In other embodiments, a control terminal 7 is also included. The control terminal 7 can be a smart terminal such as a mobile phone, tablet or computer, which can display various information such as images, motion parameters, and robotic arm operating status, and can be used by the operator to finally determine the target point and other operations.

[0033] The control flow of the robotic arm in this embodiment is as follows: Based on the location of the target, the actuator is controlled to move to a position close to the target. For example, after the robotic arm of this embodiment is moved to the vicinity of the target, the robotic arm body 1 is manually controlled to move to the position close to and facing the target.

[0034] At this time, the operator can move away from the target and send scanning commands to the robotic arm body 1 via remote communication.

[0035] In response to the scanning command, the control module 5 controls the detection module 4 to scan the target and the surrounding environment to obtain scanning information according to the task. Based on the scanned information, a three-dimensional model of the target task and its surrounding environment is constructed. More specifically, the control module 5 controls the detection module 4 to scan the target and its surrounding environment to obtain scanning information according to the task, including the following steps: The control module 5 controls the detection module 4 to scan at a first speed. Perform an initial scan of the target and its surrounding environment to obtain initial scan information; Based on the initial scan information, an initial model of the task target and its surrounding environment is constructed; In the initial model, a specific area containing the target task is selected, and the second scanning speed of the detection module 4 is controlled. A second scan is performed on the specific area to obtain second scan information, wherein the first scan speed... Greater than the second scan speed ; Based on the secondary scan information, the initial model is optimized to obtain a three-dimensional model of the target and its surrounding environment.

[0036] For example, in the initial scanning stage, the area where the target is located is scanned by LiDAR + RGB-D camera to quickly obtain area information. In this step, the robotic arm body 1 can maintain a high moving speed to improve the efficiency of the initial scanning stage.

[0037] Based on the acquired regional information, a 3D model is constructed using the TSDF Volume algorithm, Poisson reconstruction, or NDT registration to obtain a 3D model containing the operational target and its surrounding environment. The specific modeling methods are existing technologies, and those skilled in the art can refer to existing 3D scanning modeling methods for understanding; this embodiment does not limit these methods.

[0038] A specific area is manually delineated in the 3D model by selecting a box or inputting a coordinate range. The specific area refers to a small area containing the target. After the specific area is delineated, the control system controls the detection module 4 at the end of the robotic arm 1 to move within the selected specific area. At this time, the moving speed of the robotic arm 1 needs to be reduced to perform high-precision scanning, while the scanning frequency of the detection module 4 is increased to ensure that a high-precision secondary scan is completed and high-precision information of the target within the specific area is obtained.

[0039] The high-precision information obtained from the second scan is input into the aforementioned initial model, and the initial model is refined to obtain a three-dimensional model containing high-precision operation target information.

[0040] In practical applications, the execution speed of the initial and secondary scans, i.e., the scanning speed, is the main factor affecting scanning accuracy and the final scanning result. When the scanning speed is too high, it will lead to the inability to obtain scanning information that meets the accuracy requirements. When the scanning speed is too low, it will affect the overall work efficiency. This is especially fatal in large-scale operations, which may lead to a significant extension of the work period. Therefore, it is necessary to control the scanning speed to achieve a balance between scanning accuracy and efficiency, so as to complete the work tasks efficiently and reliably.

[0041] Based on actual operational experience, the applicant found that the main factors affecting scanning speed are the geometric complexity of the target and the interference level of the working environment. Therefore, in this embodiment, these two factors are considered when adjusting the aforementioned first scanning speed. Second scan speed The calculations are as follows: Construct complexity coefficients to characterize the complexity of the task objective and its surrounding environment. Specifically, by collecting data on various types of task objectives, including the geometric complexity coefficients of the task objectives and the adapted first scan speed... Second scan speed .

[0042] Establish the aforementioned complexity coefficients With the first scan speed Second scan speed The complexity-speed mapping relationship; that is, the complexity coefficients of various types of task objectives collected above. Its applicable first scan speed Second scan speed Construct a mapping database so that each complexity coefficient The first scan speed can be matched within the range. Second scan speed Specifically, the appropriate scanning speed is input based on human experience.

[0043] Based on the initial model, the complexity coefficient of the task objective is calculated. : , , , , , in, The geometric complexity coefficient of the task objective is represented. Indicates the environmental interference coefficient. and These represent the weights of the geometric complexity coefficient and the environmental disturbance coefficient, respectively. The geometric complexity evaluation value represents the objective of the task. The standard value representing the geometric complexity of the task objective; The coefficient representing the rate of change of surface curvature of the target object. The contour regularity coefficient represents the objective of the task. and These represent the weights of the surface curvature change rate coefficient and the contour regularity coefficient, respectively. Surface curvature change rate coefficient It is mainly used to reflect the density of protrusions, depressions, and edges on the target surface. Generally, the higher the density, the greater the curvature change, and the higher the complexity of the target, and vice versa. For example, the surface curvature change rate coefficient. The calculation process is as follows: The lidar point cloud data of the target is "voxelated"; The principal curvature of all points within each voxel is calculated using the curvature function and denoted as follows: And through the formula Calculate the overall curvature at each point and calculate the average curvature of all voxels. and ,but .

[0044] The contour regularity coefficient represents the regularity of the target's outline, reflecting the degree of irregularity of the target's edges (the more sharp angles and broken lines, the higher the complexity). For example, the contour regularity coefficient... The calculation process is as follows: Based on the color image from the RGB-D camera, the two-dimensional edge contour of the target is extracted using the Canny edge detection algorithm; Perform a "polygon approximation" on the edge contour to obtain the number of vertices of the approximated polyline. , Calculate the perimeter of the minimum bounding rectangle of the target contour. , Contour regularity coefficient .

[0045] coefficient of change of surface curvature Contour regularity coefficient The normalization process involves using a cube as the standard regularity body, calculating the surface curvature change rate coefficient of the cube as the standard value of the curvature change rate, and calculating the contour regularity coefficient of the cube as the standard value of the contour regularity coefficient. Based on these two standard values, the surface curvature change rate coefficient is then normalized. Contour regularity coefficient After normalization, a weighted sum is performed to obtain the complexity coefficients used to characterize the surface complexity of the task objective. .

[0046] , This represents the environmental disturbance evaluation value for the operational objective. This represents the standard value of environmental disturbance to the operational target. The obstacle density coefficient representing the environment surrounding the target operation. This represents the distance coefficient between the target object and obstacles. and These represent the weights of the obstacle density coefficient and the spacing coefficient, respectively. For example, The calculation process for the obstacle density coefficient representing the environment surrounding the target is as follows: Based on the aforementioned three-dimensional model, the radius of influence is set with the geometric center of the target as the center of the sphere. The impact on the ball.

[0047] From the lidar point cloud data, select the point clouds that affect the sphere but are not operational targets, and calculate the volume of the affected sphere. Calculate obstacle density , To influence the total number of obstacle point clouds within the sphere.

[0048] The distance coefficient between the target and obstacles is represented by the following calculation process: The point cloud data of the target is subjected to "convex hull extraction" to obtain the convex hull surface of the target (representing the outer contour boundary of the target). The point cloud data of the obstacle is also subjected to convex hull extraction to obtain the convex hull surface of the obstacle; Calculate the minimum Euclidean distance between two convex hull surfaces. If it affects the absence of obstacles inside the ball, then let .

[0049] The aforementioned obstacle density and minimum Euclidean distance are normalized: Obstacle density normalization: When calibrating for unobstructed conditions Dense obstacle environment dot cm 3 .

[0050] Minimum Euclidean distance normalization: The closer the distance, The larger the value, when the distance is greater than or equal to 200mm, .

[0051] Calculated through the aforementioned steps and environmental interference coefficient The complexity coefficient is then calculated by weighted summation. .

[0052] Based on the complexity coefficient of the stated task objective Based on the complexity-speed mapping relationship, a first scanning speed suitable for the target task is found. Second scan speed .

[0053] Specifically, the aforementioned steps establish a complexity-speed mapping relationship, and the complexity coefficient is calculated accordingly. Then, based on the complexity coefficient Find the first scan speed that matches Second scan speed The first scan speed Second scan speed The scanning speed during the target operation is designed to match the target operation.

[0054] Based on the task, optimal path planning is performed using the 3D model to obtain the optimal running path for the execution end; Based on the obtained optimal running path, control the execution end to move to the initial predetermined position; Based on the positional deviation between the target position and the initial position, the fine-tuning platform 2 is controlled to move the robotic arm body 1 to the target position, thereby moving the execution end to the target position. The robotic arm body 1 is instructed to perform a predetermined operation.

[0055] Specifically, the coordinate data of the robotic arm body 1, the fine-tuning platform 2, and the target are acquired and normalized to obtain a normalized coordinate system. Based on the normalized coordinate system, the relative coordinate relationship between the current execution terminal and the task target is calculated. Based on the relative coordinate relationship and the task, the optimal running path of the execution terminal is planned through a path planning algorithm.

[0056] For example, first define the original datum (i.e., "coordinate system source") for the three key coordinates and record their core parameters: Robotic arm body 1 end coordinates Original coordinate system: The robot arm's own "base coordinate system" (usually with the center of the robot arm's base as the origin, the X and Y axes are established along the horizontal plane, and the Z axis is in the vertical direction).

[0057] Coordinate parameters: The angles of each joint are obtained through the encoder built into the robotic arm, and the position coordinates of the end effector are calculated by combining the kinematic model of the robotic arm (such as the DH parameter method). , , ) and posture (such as Euler angles) , , Or quaternions).

[0058] Fine-tuning platform 2 walking platform coordinates Original coordinate system: The "platform base coordinate system" of fine-tuning platform 2 (with the center of the fixed base of the platform as the origin, the X and Y axes along the horizontal movement direction of the platform, and the Z axis as the vertical lifting direction).

[0059] Coordinate parameters: The position coordinates are obtained by directly reading the displacement of the walkie-camera in the X, Y, and Z axes through the encoder of the platform's servo motor. , , (The posture is usually a fixed value because the platform only performs translational motion).

[0060] Coordinates of the specific area of ​​the task target Original coordinate system: the "sensor coordinate system" of detection module 4 (with the detection center of detection module 4 (such as lidar) as the origin, and the X, Y, and Z axes defined according to the sensor installation direction).

[0061] Coordinate parameters: The position coordinates of a specific area of ​​the target (such as a grasping point) in the sensor coordinate system are obtained through secondary scanning. , , ) and posture ( , , (e.g., the direction of the normal to the target surface).

[0062] II. Defining a Globally Unified Coordinate System To eliminate deviations, a global coordinate system needs to be set as a unified reference. Usually, the base coordinate system of the fine-tuning platform 2 is chosen as the global coordinate system (reason: the fine-tuning platform 2 is the mounting base of the robotic arm, its position is fixed and its stability is high, so it can be used as the spatial reference of the entire system).

[0063] Global coordinate system parameters: Origin: Geometric center of the fixed base of fine-tuning platform 2; X-axis: The extension direction of the 2X direction motion guide rail of the fine-tuning platform; Y-axis: The extension direction of the 2Y direction motion guide rail of the fine-tuning platform (perpendicular to the X-axis). Z-axis: Perpendicular to the XY plane (vertically upward).

[0064] III. Establish the transformation relationships from each original coordinate system to the global coordinate system. The original coordinates are transformed to the global coordinate system using a coordinate transformation matrix (containing translation and rotation parameters). The transformation relationships need to be determined through calibration during system initialization and stored in control module 5.

[0065] 1. Robotic arm end-effector coordinates → Global coordinate system The robotic arm base is mounted on the walking platform of the fine-tuning platform 2, therefore the conversion is done in two steps: ① Robotic arm end coordinates (robotic arm base coordinate system) → Robotic arm base coordinates (robot arm base coordinate system origin): By back-calculating the kinematic model of the robotic arm, the coordinates of the robotic arm base in its own base coordinate system are obtained (usually (0,0,0), since the origin of the base coordinate system is the center of the base).

[0066] ② Robotic arm base coordinates → Global coordinate system: The robotic arm base is rigidly connected to the fine-tuning platform 2's walking platform, and their relative positions are fixed (after calibration, the offset is fixed). Therefore, the coordinates of the robotic arm base in the global coordinate system are: ( , , (To fine-tune the coordinates of the platform 2 walking platform).

[0067] Finally, the coordinates of the robotic arm end effector in the global coordinate system are: .

[0068] 2. Fine-tune the coordinates of the platform 2 traveling frame → global coordinate system The original coordinate system of the fine-tuning platform 2 traveling table is consistent with the global coordinate system (since the global coordinate system is defined as the platform base coordinate system), so no complex transformation is required, and its displacement in the X, Y, and Z axes can be read directly: Global coordinates = ( , , ).

[0069] 3. Coordinates of a specific area of ​​the task target → Global coordinate system Detection module 4 is installed at the end of the robotic arm, and the relative positions of the two are fixed. The conversion is carried out in three steps: ① Target coordinates (sensor coordinate system) → Robotic arm end coordinates (robot arm base coordinate system): Installation offset between sensor and robotic arm end Given the rotation angle (fixed after calibration), the following calculations are performed using the transformation matrix: The target's coordinates in the robotic arm end coordinate system = target coordinates in the sensor coordinate system + installation offset (including rotation correction).

[0070] ② Coordinates of the robotic arm end point → Coordinates of the robotic arm base (same as ① in step 1).

[0071] ③ Robotic arm base coordinates → Global coordinate system (same as step 1, step ②).

[0072] Finally, the coordinates of the specific target region in the global coordinate system are: (Complete calculation requires the use of rotation matrices).

[0073] Through the above steps, the fine-tuning platform 2, the robotic arm body 1, and the target object can be normalized to the same coordinate system (the coordinate system where the fine-tuning platform 2 is located) to facilitate subsequent calculations and processing.

[0074] Based on the established coordinate system, the relative coordinate relationship between the robotic arm end effector and the target task is calculated. Optimal path planning is then performed based on the obtained relative coordinate relationship and task information to obtain the optimal running path for the robotic arm end effector to complete the task. Specifically, this includes: Step 1: Secondary confirmation of basic data and coordinate system (preparation for path planning) Path planning requires ensuring the accuracy of all input data. This necessitates verifying the "unified coordinate system" and key coordinate data to prevent planning failures due to data discrepancies. Coordinate system validity verification: Confirm the previously completed "coordinate normalization" results (the coordinates of the robotic arm end, the coordinates of the fine-tuning platform 2 traveling table, and the coordinates of the working target are all unified to the same world coordinate system). Read the coordinate log through the control system to check for coordinate drift (such as the cumulative error of the servo motor of the fine-tuning platform 2 or the scanning noise of the detection module). If the error exceeds the threshold (such as >0.1mm), coordinate normalization needs to be performed again.

[0075] Key coordinate data extraction: Extract two types of core data from the control system database: Starting point data: The current position coordinates of the robotic arm end effector. "and attitude parameters (such as Euler angles)" or quaternion (The attitude parameter is used to ensure that the end gripping / operation posture matches the target, such as the gripper facing the work surface.) Endpoint data: "Target location coordinates" of a specific area of ​​the task objective. "and target attitude parameters" (For example, in a gripping task, the target's posture must be consistent with the opening and closing direction of the gripper.)

[0076] Job task parameter analysis: Transform abstract task assignments into quantifiable parameters, for example: Task types: Gripping (requires "approach → grab → lift"), Assembly (requires "align → insert → fit"), Inspection (requires "scan path to cover target area"); Motion constraints: maximum permissible speed (e.g., ≤50mm / s, to avoid impact), maximum acceleration (e.g., ≤10mm / s², to prevent vibration), and operational accuracy requirements (e.g., target position error ≤±0.05mm). Task priority: such as "shortest time", "shortest path", "smooth motion".

[0077] Step 2: Quantifying the relative relationship between the robotic arm end effector and the target object Based on a unified coordinate system, the relative difference between the "current state" and the "target state" of the robotic arm end effector is calculated to provide core input for path planning (which must include both positional and orientation relationships): Calculation of relative coordinates of position: Using the target position of the task objective as a reference, calculate the vector difference between the current position of the robotic arm end and the target position; The formula is as follows: relative position vector ; This vector directly reflects the "distance" the end needs to move and the "direction of movement" in the X, Y, and Z directions (e.g., ΔX=+50mm means that it needs to move 50mm along the positive X-axis).

[0078] Calculation of relative attitudes: The posture of the robotic arm end effector needs to match the target (e.g., the gripper opening direction needs to be aligned with the target contour during gripping). The posture difference is usually quantified in the following ways: Euler angles (α, β, γ, corresponding to rotation angles about the X / Y / Z axes respectively) are represented as follows: Relative attitude angle .

[0079] Step 3: Definition and Quantification of Path Planning Constraints The optimal path must satisfy three major constraints: physical feasibility, environmental safety, and operational accuracy. These constraints need to be converted into quantifiable indicators that the algorithm can recognize. Physical constraints: limits of movement of robotic arm joints (e.g., rotation range of joint 1 -90° to 90°), upper limit of joint speed / acceleration, and load limit.

[0080] Environmental constraints: obstacle avoidance requirements (such as avoiding surrounding equipment, workbench edges, and other obstacles), and work space boundary limitations.

[0081] Accuracy constraints: allowable range of target position error and allowable range of attitude error (e.g., attitude deviation during clamping ≤ ±2°).

[0082] Step 4: Selection and parameter configuration of the optimal path planning algorithm The appropriate algorithm should be selected based on the complexity of the task scenario (such as whether obstacle avoidance is required or the level of accuracy required). The core principle is to balance "planning efficiency" and "path optimality". Optional algorithms include: Joint space planning algorithm, Cartesian space planning algorithm, A Algorithms such as Dijkstra's algorithm can be used to plan the motion trajectory of the end of the robotic arm body 1 based on the relative position vector calculated above and the added constraints.

[0083] Step 5: Path Generation and Trajectory Discretization (From "Path" to "Executable Instruction") The "continuous path" generated by the algorithm needs to be converted into "discrete motion commands" that can be executed by the robotic arm control system. The core is to map the Cartesian space path to the joint space through inverse kinematics: Inverse kinematics calculation of the path The movement of the robotic arm is driven by joints, requiring the "path (X,Y,Z,α,β,γ) of the end effector in Cartesian space" to be converted into "the angle of each joint". (corresponding to a six-axis robotic arm): Input: Discrete path points in Cartesian space (usually sampled uniformly over time or distance, such as sampling one point every 0.02s); Calculation: Solve the joint angle combination corresponding to each path point using the inverse kinematics model of the robotic arm (such as the kinematic equations established by the DH parameter method). If there are multiple solutions, select the solution with the smallest joint range of motion. Verification: Ensure that the solved joint angles are within the range of "physical constraints" (i.e., do not exceed the joint limits). If they do exceed the limits, backtrack and adjust the path points.

[0084] Trajectory smoothing To avoid "impact" or "shaking" during robotic arm movement, the discrete joint angle trajectories need to be smoothed and optimized. Speed / Acceleration Planning: Using a "trapezoidal speed curve" or "S-shaped speed curve", the speed and acceleration of the joint are continuously changed during the process of "starting → constant speed → deceleration → stopping" (e.g., the acceleration increases linearly from 0 to the maximum during the starting phase to avoid instantaneous impact).

[0085] Formatted output of path instructions The smoothed joint angle trajectory is converted into "executable instructions" according to the control system's protocol format (such as Modbus, EtherCAT), including: The joint angle value corresponding to each time step (e.g., 0.01s); The speed and acceleration settings for each joint; Operation action trigger command (e.g., after reaching the target position, trigger the clamping mechanism 3 to close).

[0086] Through the above steps, the optimal running path can be planned based on the coordinates of the target position and the current position coordinates of the end of the robotic arm 1, and the optimal running path can be converted into a running command that can be recognized by the robotic arm 1. When the robotic arm 1 runs the command, it will reach the working position along the planned path.

[0087] After reaching the initial position, the control system compares the coordinate difference between the position of the end of the robotic arm body 1 and the working position, calculates the coordinate difference between the two positions in the X, Y, and Z directions, and controls the walking platform of the fine adjustment platform 2 to move according to the coordinate difference, so as to achieve secondary fine adjustment and improve the execution accuracy of the robotic arm body 1.

[0088] Specifically, the following steps are included: Obtain the coordinate information of the target location. Obtain the coordinate information of the initially determined position. ; Calculate the positional deviation between the target position and the initial position. : , Control the fine-tuning platform 2 according to the obtained position deviation move.

[0089] Furthermore, since the robotic arm of this application is mostly used in harsh environments, such as high temperatures in enclosed spaces or large vibrations caused by uneven ground, these conditions will affect the accuracy of the robotic arm body 1, resulting in a deviation between the obtained execution end coordinates and the actual coordinates. Therefore, it is necessary to calculate and compensate for the deviation. The specific methods are as follows: Calculate the motion compensation amount of the robotic arm body 1 Combined with the aforementioned motion compensation amount and positional deviation The movement of the fine-tuning platform 2 is controlled, and the motion compensation amount is... The calculation steps include: A first prediction model is constructed with ambient temperature as input and the X-axis offset, Y-axis offset, and Z-axis offset of the robotic arm body 1 as output; a second prediction model is constructed with vibration acceleration as input and the X-axis offset, Y-axis offset, and Z-axis offset of the robotic arm body 1 as output. Specifically, based on the requirements of small sample sizes, high precision, and easy deployment in industrial scenarios, Gradient Boosting Regression Tree (GBRT) or lightweight neural networks are preferentially selected as the prediction model. Taking the training process of the first prediction model as an example, multiple sets of three-axis offsets of the robotic arm body 1 under different environmental temperatures are collected as sample data. The prediction model is trained using the sample data, and after multiple iterations, a prediction model that can predict the three-axis offset of the robotic arm body 1's actuator based on environmental temperature can be obtained. The training process of the second prediction model is similar to that of the first prediction model and can be understood by referring to the process of the first prediction model, so it will not be repeated here.

[0090] The ambient temperature of the target object and the vibration acceleration of the robotic arm body 1 are collected in real time. The collected ambient temperature is input into the first prediction model to obtain the X-axis temperature error of the robotic arm body 1. Y-axis temperature error and Z-axis temperature error ; The collected vibration acceleration is input into the second prediction model to obtain the X-axis vibration error of the robotic arm body 1. Y-axis vibration error and Z-axis vibration error ; The motion compensation amount Represented as: , , in, This indicates the X-axis compensation amount. Indicates the Y-axis compensation amount. This indicates the Z-axis compensation amount.

[0091] In summary, this step can calculate the offset of the actuator in the X, Y, and Z axes under real-time ambient temperature and vibration intensity. Based on the offset, the robot arm body 1 is moved by the fine-tuning platform 2 to compensate, thereby improving the positioning accuracy of the actuator of the robot arm body 1.

[0092] In the description of this application, it should be understood that the orientation or positional relationship indicated by directional terms such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" is usually based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing this application and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the scope of protection of this application.

[0093] For those skilled in the art, various other corresponding changes and modifications can be made based on the technical solutions and concepts described above, and all such changes and modifications should fall within the protection scope of the claims of this application.

Claims

1. An automatically controlled robotic arm, characterized in that, include: The robotic arm body has an actuator for performing actions; A gripping mechanism is located at the end of the robotic arm body and is used to perform gripping actions; A detection module, located at the end of the robotic arm, is used to detect the target and the surrounding environment. A fine-tuning platform is provided, on which the robotic arm body is mounted for secondary adjustment of the position of the robotic arm body. The control module is connected to the robotic arm body, gripping mechanism, detection module and fine-tuning platform respectively; The control module controls the detection module to scan the target and surrounding environment to obtain scanning information according to the task. Based on the scanned information, a three-dimensional model of the target task and its surrounding environment is constructed. Based on the task, optimal path planning is performed using the 3D model to obtain the optimal running path for the execution end; Based on the obtained optimal running path, control the execution end to move to the initial predetermined position; Based on the positional deviation between the target position and the initial position, the fine-tuning platform is controlled to move the robotic arm body to the target position, thereby moving the execution end to the target position. The robotic arm body is instructed to perform a predetermined operation.

2. The automatically controlled robotic arm according to claim 1, characterized in that, The gripping mechanism includes a pneumatic gripper or an electric gripper; the detection module includes one or more of an infrared detector, an ultrasonic detector, a lidar, and a 3D scanner.

3. The automatically controlled robotic arm according to claim 1, characterized in that, The fine-tuning platform includes an X-axis walking module, a Y-axis walking module, a Z-axis walking module, a first drive component, a second drive component, and a third drive component. The first drive component is linked with the X-axis walking module, the second drive component is linked with the Y-axis walking module, and the third drive component is linked with the Z-axis walking module.

4. The automatically controlled robotic arm according to claim 3, characterized in that, It also includes a displacement sensor, which is disposed at the fine-tuning platform to detect the displacement of the fine-tuning platform.

5. The automatically controlled robotic arm according to claim 1, characterized in that, It also includes a control terminal, which is signal-connected to the control module.

6. The automatically controlled robotic arm according to claim 1, characterized in that, The control module controls the detection module to scan the target and its surrounding environment to obtain scanning information according to the task, including the following steps: The control module controls the detection module to scan at a first speed. The target and its surrounding environment are initially scanned to obtain initial scan information. Based on the initial scan information, an initial model of the task target and its surrounding environment is constructed; In the initial model, a specific area containing the target task is selected, and the second scanning speed of the detection module is controlled. A second scan is performed on the specific area to obtain second scan information, wherein the first scan speed... Greater than the second scan speed ; Based on the secondary scan information, the initial model is optimized to obtain a three-dimensional model of the target and its surrounding environment.

7. The automatically controlled robotic arm according to claim 6, characterized in that, The control module controls the detection module to scan the target and its surrounding environment according to the task, and obtains the following scanning information: Construct complexity coefficients to characterize the complexity of the task objective and its surrounding environment. , Establish the aforementioned complexity coefficients With the first scan speed Second scan speed The complexity-speed mapping relationship; Based on the initial model, the complexity coefficient of the task objective is calculated. : ， ， ， ， ， in, The geometric complexity coefficient of the task objective is represented. Indicates the environmental interference coefficient. and These represent the weights of the geometric complexity coefficient and the environmental disturbance coefficient, respectively. The geometric complexity evaluation value represents the objective of the task. The standard value representing the geometric complexity of the task objective; The coefficient representing the rate of change of surface curvature of the target object. The contour regularity coefficient represents the objective of the task. and These represent the weights of the surface curvature change rate coefficient and the contour regularity coefficient, respectively. This represents the environmental disturbance evaluation value for the operational objective. This represents the standard value of environmental disturbance to the operational target. The obstacle density coefficient represents the obstacle density coefficient in the environment surrounding the target operation. This represents the distance coefficient between the target object and obstacles. and These represent the weights of the obstacle density coefficient and the spacing coefficient, respectively. Based on the complexity coefficient of the stated task objective Based on the complexity-speed mapping relationship, a first scanning speed suitable for the target task is found. Second scan speed .

8. The automatically controlled robotic arm according to claim 1, characterized in that, Based on the task and the 3D model, the optimal path planning for obtaining the optimal running path of the execution terminal includes the following steps: The coordinate data of the robotic arm body, the fine-tuning platform, and the target are acquired and normalized to obtain a normalized coordinate system. Based on the normalized coordinate system, the relative coordinate relationship between the current execution terminal and the task target is calculated. Based on the relative coordinate relationship and the task, the optimal running path of the execution terminal is planned through a path planning algorithm.

9. The automatically controlled robotic arm according to claim 1, characterized in that, Based on the positional deviation between the target position and the initial position, controlling the fine-tuning platform to move the robotic arm body to the target position includes the following steps: Obtain the coordinate information of the target location. Obtain the coordinate information of the initially determined position. ; Calculate the positional deviation between the target position and the initial position. : ， Control the fine-tuning platform according to the obtained position deviation move.

10. The automatically controlled robotic arm according to claim 9, characterized in that, Controlling the movement of the fine-tuning platform also includes: calculating the motion compensation amount of the robotic arm body. Combined with the aforementioned motion compensation amount and positional deviation Controlling the movement of the fine-tuning platform, the motion compensation amount The calculation steps include: A first prediction model is constructed with ambient temperature as input and the X-axis, Y-axis, and Z-axis offsets of the robotic arm body as outputs; a second prediction model is constructed with vibration acceleration as input and the X-axis, Y-axis, and Z-axis offsets of the robotic arm body as outputs. The ambient temperature of the target object and the vibration acceleration of the robotic arm body are collected in real time. The collected ambient temperature is input into the first prediction model to obtain the X-axis temperature error of the robotic arm body. Y-axis temperature error and Z-axis temperature error ; The collected vibration acceleration is input into the second prediction model to obtain the X-axis vibration error of the robotic arm body. Y-axis vibration error and Z-axis vibration error ; The motion compensation amount Represented as: ， ， in, This indicates the X-axis compensation amount. Indicates the Y-axis compensation amount. This indicates the Z-axis compensation amount.

Images